Testing for Tolerance in a Box Experiment in Wild Vervet Monkeys - M.AungKyaw - 2022-2023
juin 19, 2024 19:24
Introduction
0.Opening the data
Loading data
- First I downloaded the knitr package to create outputs as html, pdf or word files when knitting my r markdown file. I also loaded the pander package for better presentation
- The dplyr package was installed for better manipulation of the data as filtering or creating new variables and lubridate for a better manipulation of dates and time
- Then, I installed the readxl package to import my dataset which is called Box Experiments.xls
- This dataset contains information related to my master thesis project. I used cyber tracker in order to record the behaviors of dyads of Vervet monkeys in a box experiment on tolerance from September 2022 to September 2023
1.Explore the data
Description of the initial datset - “Boxex”
## Glimpse of the Box Experiment dataset:## Rows: 2,795
## Columns: 20
## $ Date <dttm> 2022-09-27, 2022-09-27, 2022-09-27, 2022-09-27,…
## $ Time <dttm> 1899-12-31 09:47:50, 1899-12-31 09:50:07, 1899-…
## $ Data <chr> "Box Experiment", "Box Experiment", "Box Experim…
## $ Group <chr> "Baie Dankie", "Baie Dankie", "Baie Dankie", "Ba…
## $ GPSS <chr> "-28.010549999999999", "-28.010549999999999", "-…
## $ GPSE <chr> "31.191050000000001", "31.191050000000001", "31.…
## $ MaleID <chr> "Nge", "Nge", "Nge", "Nge", "Nge", "Nge", "Nge",…
## $ FemaleID <chr> "Oerw", "Oerw", "Oerw", "Oerw", "Oerw", "Oerw", …
## $ `Male placement corn` <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
## $ MaleCorn <dbl> 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, …
## $ FemaleCorn <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
## $ DyadDistance <chr> "2m", "2m", "1m", "1m", "0m", "0m", "0m", "0m", …
## $ DyadResponse <chr> "Tolerance", "Tolerance", "Tolerance", "Toleranc…
## $ OtherResponse <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
## $ Audience <chr> "Obse; Oup; Sirk", "Obse; Oup; Sirk", "Oup; Sirk…
## $ IDIndividual1 <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
## $ IntruderID <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, "Sey…
## $ Remarks <chr> NA, NA, "Nge box did not open because of the bat…
## $ Observers <chr> "Josefien; Michael; Ona; Zonke", "Josefien; Mich…
## $ DeviceId <chr> "{7A4E6639-7387-7648-88EC-7FD27A0F258A}", "{7A4E…I am now using the View function to have a sight on the entire dataset and glimpse to display a summary of my dataset
I have 20 variables (here columns) and 2795 trials (here rows)
I will now make a brief summary of each variables and their use before creating a new dataframe (df) with my variables of interest that I will call Bex
The highlighted variables are the ones I will use for Bex. I will then clean the data before heading to the statistical analysis and the interpretation of the results
Variables of Boxex
Date : “Date” is in a POSIXct format which is appropriate for the display of time
- I want to use the date to know how many sessions have been done with each dyads in my experiment.
- I will create a variable called Session where 1 session = 1 day
- The data has values from the 14th of September 2022 until the 13th of September 2023
- I may consider separating the 12 months of data in 4 seasons to make a preliminary check of a potential effect of seasonality. Nevertheless since we did not use any tools to measure the weather, temperature, humidity or food availability (also related to seasonality and weather). Categorizing my data in 4 without having further data would then be quite arbitrary. If I end up doing it in my report, it will be done without any intention to include it in my scientific analysis nor my scientific report .
Time : “Time is coded” in a POSIXct format
- I do not plan to use this variable but we can see that “Time” has the correct hours displayed with a date which is incorrect.
- (In the case I wanted to observe when the trials occurred during the day as time may have an influence on their behavior (Isbell & Young 1993)) I would need to correct the incorrect display of the date in the dataset.
- This variable could also be useful to see when the seasonal effect took place as we only went in the morning during summer because of the heat while we went later and for longer times in the field to do the box experiment in winter
- For now, the values in “Time” are all on the same (wrong) day which is the 31st of December
- Note: I first did not intend to keep Time in Bex but I needed this variable to see the order of the trails within a day. I finally decided to keep it.
Data : chr “Data” is coded as character
- It describes the type of data being recorded in the software cybertracker. We installed the software on tablets to record the different behaviors of vervet monkeys in our research center
- In our case, my data was recorded in cybertracker as Box Experiment as we created a form specifically for this experiment
- For this reason we can remove this column since the information it contains is unecessary and redundant
Group : chr The data is coded in r as a character
- It describes the group of monkey in which we did the trial
- I will keep this column to see the amount of trials that we did in the 3 group of monkeys which are Baie-Dankie (BD), Ankhase (AK), and Noha (NH)
GPSS : num “GPSS” is coded as numerical
- It gives the south coordinates in which we started the experiment
- I do not plan to use coordinates nor look at locations so I will remove this acolumn
GPSE : num “GPSE” is coded in as numerical
- It gives the east coordinates in which we started the experiment
- I do not plan to use coordinates nor look at locations so I will remove this column
MaleID : chr “MaleID” is coded as character
- It indicates the name of the male involved in the trial
- I plan to use this to see how factors related to the individual may influence the experiment (age, sex, rank)
- It will also help me see which behaviour was displayed by each individuals (here males)
FemaleID : chr “FemaleID” is coded as character
- It indicates the name of the female involved in the trial
- I plan to use this variable in the same way as “Male ID”
- It will also help me see which behaviour was displayed by each individuals (here females)
Male placement corn: dbl “Male placement corn is coded in r as double
It gives the amount of corn given to the male of the dyad before the trials
Within a session it happened that we gave more placement corn to attract the monkeys again to the boxes. This lead to an update of the number in the same session. The number found at the end of the session is the total placement corn an individual has received
I will fuse this column with male corn as the data has been separated between these two variables. This is due to a mistake when creating the original box experiment form in cybertracker
This variable could be related to the level of motivation of a monkey but as it is not directly related to my hypothesis I may not use this column. I will re-consider the use of this column later on
In regards of this possibility I will change the format of the variable to numerical
MaleCorn : dbl “MaleCorn” is coded in r as double
- It gives the same information as in male placement corn
- I will import the values from “male placement corn” into this one
- I will change the format of the variable to numerical
FemaleCorn : dbl The data is coded in r as double
- It gives the amount of corn given to the female of the dyad before the trials
- It works in the same way as “male placement corn”/“MaleCORN”
- I will change the format of the variable to numerical
DyadDistance : chr The data is coded in r as character
- It gives the distance for each trial that we have done with the dyads.
- The trial number 1 for each dyad was at 5 meters.
- The maximum was around 10 m while the minimum is 0
- We will have to remove the “m” for meters in order to have a numerical variable instead of character
- Also, since the very first trials per dyad can be considered as a kind of learning phase, i may remove the 15 first trials that were made for each dyad
DyadResponse : chr The data is coded in r as character
- It indicated which behaviour was produced by the dyad’s during each trial
- The different behaviours were: Distracted, Female aggress male, Male aggress female, Intrusion, Loosing interest, Not approaching, Tolerance and Other
- I will change the columns associated to each behavior (i.e. Response) of DyadResponse into dichotomic variables in order to see the frequency of each behaviour
- This will allow me to see which behavior occurred more ,and behavioural differences could be found between dyads
- As multiple response could occur within the same trial, multiple behaviors can be found in a single cell. I will create a hierarchy to reduce the amount of behaviors assigned to each trial (if there is more than one). This will also be complemented with the information found in the column remarks
- correct any mistakes (ex. if tolerance and aggression are together aggression>tolerance)
- assign as few labels per trial
- get a better View and understanding of the data and the most common behaviours produced by each dyad
- create variables that can complement the behaviour found (ex. not approaching + looks at partner would be looks at partner + a new variable called hesistant to see when the did not come but look at the other individual / )
- Projection of the hierarchy (changes will be made)
Create a table with each combination existing
Decide what is more important
Ex:
- Aggression > Tolerance
- Tolerance > Not approaching -> Create a variable called hesistant in addtion to the tolerance count to see frequency of tolerance behaviour that happened after > 1min
- Tolerance > Loosing interest
- Tolerance > Intrusion
- Not approaching = looking box but not coming while Loosing interest = not paying attention to the box
- Intrusion > Loosing interest
- Intrusion > Not approaching
- Not approaching > Looks at partner
- We can code every look at partner as no approaching and keep the count of looks at partner as additional information
- Not approaching >?> Loosing interest ? !!
- Define distracted
- Not approaching > Distracted
- Aggression > Not approaching
- Other > Look case by case and categorize depending of behavior
- Remarks may be used for the same reason
OtherResponse : chr “The data”OtherResponse” is coded as character
- It describes any behaviour that is different from the ones found in Dyad Response (meaning ≠ tolerance, aggression, intrusion, loosing interest, not approaching, distracted, looks at partner that where categorized as other)
- I will have to look at every OtherResponse and rename each entry in one of the response already if existing. I will proceed case by case.
- If I want to do an intermediate manipulation I may rename every NA in “OtherResponse” into Response to see the amount of case to treat and how many occurrences seem to not fit in the categories of “DyadResponse”
Audience : chr “Audience” is in r as character
- It gives the names of the individuals in the audience
- I would like to use it to see the amount of audience (big vs small) and the dominance level of the audience (high vs low)
- I will create a variable called NAudience to see hoy many individuals are in the audience for each trial
- After calculating the elo ratings of the individuals using another dataset (Life history), I will create a dichotomic variable called RankAudience to see effects related to rank with the effect of audience
IDIndividual1 : chr “IDIndividual1” is coded in r as character
- It gives the names of the individuals that did not approach, showed aggression, got distracted or lost interest during a trial
- I will have to look at it to see how often these behaviors occurred
- I will consider how to use this variable during the cleaning of the data
IntruderID : chr “IndtruderID” is coded as character
- It gives the name of the individual that intruded the experiment during a trial
- Intrusion could mean, invade the space of the experiment and interact with one of our individual, steal the food, show agnostic behavior, stand in very close proximity of the dyad’s individuals
Remarks : chr The data is coded in r as character
- It gives either additional information concerning the experiment when unusual behaviors occurred , mistakes that needed to be corrected or details that we wanted to record in case we would need them
Observers :chr The data is coded in r as character
- It gives the names of the observers during the experiment
- We will not use this data as we do not look at the effect that an experimenter would have on the monkeys
- (Should I still look at an effect of the amount of experimenter?…maybe better for detailled analysis of our study)
DeviceID :chr “The data”DeviceID” is coded in r as character
- It gives the name of the device/tablet used to record the data during the experiment
- We will not use this data either
2. Treating missing data
2.1. Creating a new dataframe - Bex
Since I do not want to work with the whole dataset, I’m gonna select the variables of interest using the function select
I will keep Time, Date, Group, MaleID, FemaleID, MaleCorn, Male placement corn, FemaleCorn, DyadDistance, DyadResponse, OtherResponse, Audience, IDIndividual1, IntruderID, Remarks
## Rows: 2,795
## Columns: 15
## $ Time <dttm> 1899-12-31 09:47:50, 1899-12-31 09:50:07, 1899-…
## $ Date <dttm> 2022-09-27, 2022-09-27, 2022-09-27, 2022-09-27,…
## $ Group <chr> "Baie Dankie", "Baie Dankie", "Baie Dankie", "Ba…
## $ MaleID <chr> "Nge", "Nge", "Nge", "Nge", "Nge", "Nge", "Nge",…
## $ FemaleID <chr> "Oerw", "Oerw", "Oerw", "Oerw", "Oerw", "Oerw", …
## $ MaleCorn <dbl> 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, …
## $ `Male placement corn` <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
## $ FemaleCorn <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
## $ DyadDistance <chr> "2m", "2m", "1m", "1m", "0m", "0m", "0m", "0m", …
## $ DyadResponse <chr> "Tolerance", "Tolerance", "Tolerance", "Toleranc…
## $ OtherResponse <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
## $ Audience <chr> "Obse; Oup; Sirk", "Obse; Oup; Sirk", "Oup; Sirk…
## $ IDIndividual1 <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
## $ IntruderID <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, "Sey…
## $ Remarks <chr> NA, NA, "Nge box did not open because of the bat…2.1.1 Merging Male placement corn and MaleCorn
- I want to process all the missing data in Bex. But before, I will merge the column MaleCorn and Male placement corn as the data of both columns is supposed to be together under “MaleCorn”
- Looking manually in the Bex table it seems that very few data is in MaleCorn while most of it seems to be in Male placement corn
- Every time there is a missing value in Male placement corn we can see a value in Male Corn, I will then create a new variable MaleCorn where every time that there is NA in male placement corn the value will be taken in MaleCornOld (previous malecorn). If there is no NA it will take the value of ´Male placement corn´
- I will first rename MaleCorn to MaleCornOld, then check the amount of NA’s and then merge “MaleCornOld” and “male placement corn” into the new variable “MaleCorn”
## Number of rows with common NAs in MaleCornOld and 'Male placement corn': 1499
## Number of occurrences of 0 in MaleCorn: 1499
## Number of remaining NA values in MaleCorn: 0I have found 1499 NA in common between MaleCornOld and ‘male placement corn’, 1609 NA in Male placement corn and 2685 in MaleCorn old
For the merge of MaleCornOld and Male placement corn, I used different conditions: 1.In this code, a new variable MaleCorn is created. If there is a missing value in Male placement corn, it takes the corresponding value from MaleCornOld; otherwise, it takes the value from Male placementcorn. 2.If there are no value in both MaleCornOld and Male placement corn (NA,NA) for a given row, I would like the code to display 0 as it means that no placement was given
In this way, I should not loose any data, minimize the mistakes and already transform the NA’s of this variable into a number which will remove the remaining NA’s which are meant to be 0
After the merge I found that there were no NA’s remaining in the “New” Male Corn and that 1499 0’s where found in the column which corresponds to the amount of common NA’s found previously between the “Old” Male Corn and male placement corn
2.1.2 Cleaning FemaleCorn
## Number of remaining NA values in FemaleCorn: 02.2 Cleaning variables with missing data
Now in order to see where are located the missing points in the data, I’m going to print the variables with and without NA’s
The function sapply is used to apply the function sum for NA’s to each column of the data frame, so each variable
## Variables with Missing Data:| x | |
|---|---|
| MaleID | 19 |
| FemaleID | 60 |
| DyadDistance | 33 |
| DyadResponse | 47 |
| OtherResponse | 2758 |
| Audience | 924 |
| IDIndividual1 | 2143 |
| IntruderID | 2737 |
| Remarks | 2181 |
## Variables with No Missing Data:| x | |
|---|---|
| Time | 0 |
| Date | 0 |
| Group | 0 |
| FemaleCorn | 0 |
| MaleCorn | 0 |
We can see that out of the 14 variables we have in Bex we have 9 variables with missing data which are Male ID, Female ID, DyadDistance, DyadResponse, OtherResponse, Audience, IDIndividual1, IntruderID, Remarks: I will proceed to clean these variables one by one
MaleID 19
FemaleID 60
DyadDistance 33
DyadResponse 47
OtherResponse 2758
Audience 924
ID Individual1 2143
IntruderID 2737
Remarks 2181
Before making treating the NA’s in the dataset I will make a backup of the data at this point:
2.3 Treating variables with missing data
2.3.1 Cleaning “Remarks” - (2181 NA’s)
Since most of the time we did not have any remarks it is understandable that this variable contains 2181 NA’s out of 2795 rows
I will first transform every missing data in the column Remark into No Remarks and then check that the amount of “No remarks” found
After the changes we can effectively see that we have 2181 “No Remarks” and we have no missing data left in that column, I will treat this column by hand once all the NA’s have been removed from the dataset
## Number of 'No Remarks' in the 'Remarks' column: 2181##
## No Remarks Remarks
## 2181 6142.3.2 Cleaning “Intruder ID” - (2737 NA’s)
- Intruder ID is a variable that contains the name of the individuals that made and intrusion during a trial.
- If more than one individual intruded, his name may be in the comments, which I will check when treating the data from this column
- Because nothing was entered when there was no intrusion, I will replace every NA’s by No Intrusion
- Also, I will use a function to create a new dichotomic variable called Intrusion. Every time there is a value in IntruderID, it should display 1 (Yes), if not a 0 (No intrusion)
## Number of 'No Intrusion' in the 'Intruder ID' column after replacement: 2737- We previously had 2737 NA’s in IntruderID while now we have the same amount of occurrences of IntruderID which shis that the transformation went as intended
2.3.3 Cleaning “IDIndividual1” - (2143 NA’s)
- IDIndividual1 is meant to report the name of the individual that did a behavior such as not approach, show aggression or loose interest during a trial
- I will now replace every NA in this column by No individual and print the amount of NA’s left and the amount of changes made
## Number of NAs replaced in IDIndividual1: 2143
## Number of remaining NA values in IDIndividual1: 02.6 Cleaning “Audience” - (924 NA’S)
- Audience is made to report every name of individuals around our dyad during a given trial
- I will replace every NA by No audience as no entry means the absence of other individuals around
- I will also create a new variable called “Amount audience” that will have to tell me how many individuals are found in the column Audience
## Number of changes made in 'Audience': 924## Remaining NA values in 'Audience': 02.7 Cleaning “OtherResponse” - (2758 NA’S)
## Number of changes made in 'OtherResponse': 2758## Remaining NA values in 'OtherResponse': 02.8 Cleaning of “Time”
- Since the reading of the data is more complicated without the time, which was usefull to know which trial was before or after, I changed the code made for Bex and added Time in the dataframe. Since I will need it for the cleaning of Dyaddistance, I will now extract the time from the date. Even if the date is wrong as seen in the first output, the time is correct. As in the second output, only the time has been kept
## [1] "1899-12-31 09:47:50 UTC" "1899-12-31 09:50:07 UTC"
## [3] "1899-12-31 09:53:11 UTC" "1899-12-31 09:54:28 UTC"
## [5] "1899-12-31 09:55:19 UTC" "1899-12-31 09:56:56 UTC"## [1] "09:47:50" "09:50:07" "09:53:11" "09:54:28" "09:55:19" "09:56:56"2.9 Cleaning DyadDistance
- Before looking at the NA’s of Dyaddistance I will remove the “m” that is in front of every number to have a numerical variable
- Then I will look at the location of the NA’s in the data to treat them case by case.
## Warning: NAs introduits lors de la conversion automatique## # A tibble: 69 × 16
## Time Date Group MaleID FemaleID FemaleCorn DyadDistance
## <chr> <dttm> <chr> <chr> <chr> <dbl> <dbl>
## 1 12:09:34 2022-09-27 00:00:00 Baie Da… Xia Piep 7 NA
## 2 12:13:28 2022-09-27 00:00:00 Baie Da… Xia Piep 7 NA
## 3 16:02:32 2022-09-15 00:00:00 Ankhase Sho Ginq 6 NA
## 4 10:46:33 2023-08-17 00:00:00 Baie Da… Xia Piep 0 NA
## 5 09:30:17 2023-07-29 00:00:00 Baie Da… Xin Ouli 0 NA
## 6 12:08:51 2023-07-11 00:00:00 Baie Da… Xia Piep 0 NA
## 7 13:30:07 2023-06-29 00:00:00 Baie Da… Sey Sirk 0 NA
## 8 09:54:24 2023-06-27 00:00:00 Ankhase Sho Ginq 0 NA
## 9 10:13:56 2023-06-23 00:00:00 Ankhase Sho Ginq 0 NA
## 10 09:39:04 2023-06-15 00:00:00 Ankhase Sho Ginq 2 NA
## # ℹ 59 more rows
## # ℹ 9 more variables: DyadResponse <chr>, OtherResponse <chr>, Audience <chr>,
## # IDIndividual1 <chr>, IntruderID <chr>, Remarks <chr>, MaleCorn <dbl>,
## # Intrusion <dbl>, AmountAudience <dbl>
## Number of NA values in DyadDistance column (using second approach): 69
## Rows with NA values in DyadDistance column: 24, 27, 95, 492, 744, 971, 1113, 1130, 1164, 1261, 1341, 1396, 1491, 1583, 1683, 1693, 1717, 1718, 1719, 1724, 1725, 1739, 1755, 1756, 1757, 1764, 1779, 1782, 1792, 1799, 1800, 1840, 1841, 1868, 1869, 1888, 1891, 1892, 1896, 1911, 1912, 1915, 1918, 1919, 1952, 1953, 1958, 1980, 1981, 1984, 1986, 1996, 2000, 2009, 2054, 2104, 2105, 2191, 2233, 2234, 2287, 2437, 2569, 2579, 2580, 2643, 2676, 2709, 2729We have 69 missing values in DyadDistance. I will look at each row in it’s context as the actual distance of the box was always dependent of the previous trials. I will start with the bigger number as for now the oldest trial is at the last row while the closest one is in row 1.
- If tolerance was achieved twice in a row = <1m
- If aggression (male agress female or female agress male), not approaching,or loosing interest occured = >1m
- If distracted or intrusion occured = same distance
- 24 - In trial23 (0m) there was aggression then at trial24 (1m) there was tolerance. The 24th trial is supposed to be at 1m
- 27 - In trial25 (1m) there was not approaching then at trial26 (2m) there was tolerance.The 25th trial is supposed to be at 2m
- 95 - In trial93 (2m ) there was male agress female then at trial 94 (3m) there was not approaching. The 95th trial is supposed to be at 4m
- 492 - In trial490 (0m) there was tolerance then at trial491 (0m) there was tolerance. The 492nd trial is supposed to be at 0m
- 744 - In trial742 (3m) there was aggression then at trial743 (4m) there was tolerance. The 744th trial is supposed to be at 4m
- 971 - In trial969 (0m) there was tolerance then at trial970 (0m) there was tolerance. The 971st trial is supposed to be at 0m
- 1113 - In trial1111 (2m) there was tolerance then at trial1112 (0m) there was tolerance. The 1113th trial is supposed to be at 0m
- 1130 - In trial1128 there was another dyad so we can not use this cell. Then at trial1129 (3m) there was not approaching. Nevertheless, we don’t have any DyadResponse, i will thus delete this row
- 1164 - In trial1162 (3m) there was not approaching then at trial1163 (3m) there was not approaching. The 1164th trial is supposed to be at 4m
- 1261 - The two preivous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1341 - In trial1339 (0m) there was tolerance then at trial1340 (0m) there was tolerance. The 1341st trial is supposed to be at 0m
- 1396 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1491 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1583 - In trial1581 (2m) there was not approaching and intrusion then at trial1582 (2m) there was not approaching. The 1583rd trial is supposed to be at 3m
- 1683 - One trial only was made with tolerance (2m) but since there are no DyadResponse I will delete this row
- 1693 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1717 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1718 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1719 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1724 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1725 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1739 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1755 - since there are no DyadResponse I will delete this row
- 1756 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1757 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1764 - Since there are no DyadResponse I will delete this row
- 1779 - It seems like it was the first trial of the Dyad Pom Xian, if so, the distance has to be 5m
- 1782 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1792 - Trial1791 was intrusion (4m) so this trial should be at 4m
- 1799 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1800 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1840 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1841 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1868 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1869 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1888 - he two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1891 - Since there are no DyadResponse I will delete this row
- 1892 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1896 - Since there are no DyadResponse I will delete this row
- 1911 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1912 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1915 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1918 - In trial1916 (4m) there was tolerance then at trial1917 (4m) there was not loosing interest The 1918th trial is supposed to be at 4m
- 1919 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1952 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1953 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1958 - In trial1956 (2m) there was tolerance then at trial1957 (2m) there was distracted. The 1958th trial is supposed to be at 2m
- 1980 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1981 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1984 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1986 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 1996 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 2000 - In trial1997 and 1999 (5m) there was tolerance then at trial1999 (5m) there was intrusion. The 2000th trial is supposed to be at 4m
- 2009 - Since there are no DyadResponse I will delete this row
- 2054 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 2104 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 2105 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 2191 - In trial2189 (1m) there was not approaching then at trial2190 (2m) there was not approaching. The 2191st trial is supposed to be at 3m
- 2233 - In trial2231 (3m) there was not approaching then at trial2232 (4m) there was not approaching. The 2233rd trial is supposed to be at 5m
- 2234 - The trial did not happen because they where not at the right distance. I will thus delete this row
- 2287 - Since there are no DyadResponse I will delete this row
- 2437 - Since there are no DyadResponse I will delete this row
- 2569 - In trial2567 (1m) there was tolerance then at trial2568 (1m) there was tolerance. The 2569th trial is supposed to be at (0m)
- 2579 - In trial2577 (1m) there was tolerance then at trial2578 (0m) there was not approaching. The 2579th trial is supposed to be at 1m
- 2580 - The two previous trials were made with another Dyad. Also DyadResponse is not available. I will thus delete this row
- 2643 - Since there are no DyadResponse I will delete this row
- 2676 - In trial2674 (1m) there was tolerance then at trial2675 (0m) there was tolerance. The 2676th trial is supposed to be at 0m
- 2709 - In trial2707 (2m) there was tolerance then at trial2708 (2m) there was tolerance. The 2709th trial is supposed to be at 1m
- 2729 - In trial2727 (3m) there wastolerance then at trial2728 (2m) there was tolerance. The 2729th trial is supposed to be at 2m
Now that I have looked at each missing line and saw which ones to keep, I decided to create a new variable called Distance. I will also to create a new variable called No trial.
For the variable Distance I will replace each row where there was missing data with a value and I will delete the ones where no values could be assigned. This will allow me to have no missing data and find a number to each trial that has been done
Before making the changes i’m gonna make a backup called BackupbeforeDistanceNA
## Number of NA's in DyadDistance after replacements and deletions: 1
## Data size after deletions: 2748## Row index with NA in DyadDistance: 1925*It seems that there is still the row 1925 with an NA in DyadDistance
- 1925 - In trial1923 (2m) there was distracted then at trial1924 (2m) there was tolerance. The 2725th trial is supposed to be at 2m
## Row index with NA in DyadDistance:*In this modification, I added a check to see if the columns Dyadistance and Distance already exist in your dataframe (Bex). If they do, it prints a message saying that the modification has already been applied, and no changes are made. If they don’t exist, it proceeds with the modifications. This way, running the code multiple times won’t cause redundant changes.
- 24 values were inserted in Distance to replace the NA’s where the distance could be found by looking at the previous rows. The 46 remaining NA’s were then removed from Distance leaving 0 remaining NA in the variable “Distance”
2.10 Cleaning Female and Male ID
Before cleaning Female and Male ID, here is a list of every dyad of the box experiment and their respective groups. This will help us find the missing names when only one individual is missing out of the duo (either male or female):
Sirk & Sey - BD
Ouli & Xin - BD
Piep & Xia - BD
Oerw & Nge - BD
Oort & Kom - BD
Ginq & Sho - AK
Ndaw & Buk - Ak
Xian & Pom - AK
Guat & Pom - Ak
Note that the 4 letter codes correspond to the femaleID, the 3 letter codes to the males ID and the 2 letter codes to the group name of the monkeys
I need to check where are the NA’s in both FemaleID and Male ID by looking at the rows where data is missing. Since every trial was made with a Dyad and never with an single individual, treating these two columns together makes more sense. If both individuals are missing I may have to delete the row.
## Row numbers with missing values in FemaleID: 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1884 1885 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629## Number of missing values in FemaleID: 59## Row numbers with missing values in MaleID: 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710## Number of missing values in MaleID: 18## Number of rows with missing values in both FemaleID and MaleID: 18## Row numbers with missing values in both FemaleID and MaleID: 1693, 1694, 1695, 1696, 1697, 1698, 1699, 1700, 1701, 1702, 1703, 1704, 1705, 1706, 1707, 1708, 1709, 1710## Number of missing values in FemaleID not in MaleID: 41## Row numbers with missing values in FemaleID not in MaleID: 865, 866, 867, 868, 869, 870, 871, 872, 873, 874, 875, 876, 877, 878, 879, 1808, 1809, 1810, 1811, 1812, 1813, 1814, 1815, 1816, 1817, 1818, 1819, 1820, 1884, 1885, 2619, 2620, 2621, 2622, 2623, 2624, 2625, 2626, 2627, 2628, 2629FemaleID has 41 NA’s while they are 18 NA’s in Male ID
In these missing data, we have 18 NA’s that are in common between FemaleID and MaleID which represents the totality of the missing values in MaleID
All the missing data in MaleID are found in consecutive rows, from row 1693 to row 1710 and are from the group Noha (NH) on the 19th of april 2023. We can also see that trials had bee made in the same day, and looking at the time of the experiment, the previous trials made and the audience we can see that these NA’s in female and male ID we can asses that the individuals involved were Xian for the female ID and Pom for the MaleID. I will thus replace these values using a condtion. These NA’s in Noha (Trial 1693 to 1710) are the only NA’s that MaleID has and are the only NA’s of female ID in Noha. I will thus replace every NA of MaleID NA in Noha with Pom and every Female ID NA in Noha with Xian
## Number of remaining NA values in MaleID after replacement: 0## Number of remaining NA values in FemaleID after replacement: 41## Number of rows with missing values in both MaleID and FemaleID after replacement: 0In order to clean FemaleID, I will use the data from the now complete MaleID. I will use conditions stating that depending which name is found in MaleID when there is an NA in FemaleID, a certain name will have to replace the NA in female ID
Before automating the process I will check manually the data to see if they are any exceptions or mistakes
## Rows with missing values in FemaleID:## # A tibble: 41 × 16
## Time Date Group MaleID FemaleID FemaleCorn DyadDistance
## <chr> <dttm> <chr> <chr> <chr> <dbl> <dbl>
## 1 09:31:55 2023-07-22 00:00:00 Ankhase Buk <NA> 7 1
## 2 09:33:14 2023-07-22 00:00:00 Ankhase Buk <NA> 7 1
## 3 09:34:07 2023-07-22 00:00:00 Ankhase Buk <NA> 7 0
## 4 09:34:51 2023-07-22 00:00:00 Ankhase Buk <NA> 7 0
## 5 09:36:59 2023-07-22 00:00:00 Ankhase Buk <NA> 7 0
## 6 09:38:13 2023-07-22 00:00:00 Ankhase Buk <NA> 7 1
## 7 09:39:26 2023-07-22 00:00:00 Ankhase Buk <NA> 7 0
## 8 09:41:11 2023-07-22 00:00:00 Ankhase Buk <NA> 0 0
## 9 09:42:17 2023-07-22 00:00:00 Ankhase Buk <NA> 0 0
## 10 09:44:06 2023-07-22 00:00:00 Ankhase Buk <NA> 0 1
## # ℹ 31 more rows
## # ℹ 9 more variables: DyadResponse <chr>, OtherResponse <chr>, Audience <chr>,
## # IDIndividual1 <chr>, IntruderID <chr>, Remarks <chr>, MaleCorn <dbl>,
## # Intrusion <dbl>, AmountAudience <dbl>If there is NA in femaleID, we will replace the value with - Sirk if MaleID is Sey - Ouli if MaleID is Xin - Piep if MaleID is Xia - Oerw if MaleID is Nge - Oort if MaleID is Kom - Ginq if MaleID is Sho - Ndaw if MaleID is Buk
## # A tibble: 20 × 3
## MaleID FemaleID Count
## <chr> <chr> <int>
## 1 Xia Piep 576
## 2 Sey Sirk 557
## 3 Kom Oort 338
## 4 Sho Ginq 278
## 5 Pom Xian 259
## 6 Buk Ndaw 245
## 7 Xin Ouli 159
## 8 Nge Oerw 153
## 9 Piep Xia 35
## 10 Oort Kom 29
## 11 Ouli Xin 27
## 12 Oerw Nge 19
## 13 Sirk Sey 17
## 14 Buk <NA> 15
## 15 Sey <NA> 13
## 16 Nge <NA> 11
## 17 Buk Ginq 6
## 18 Pom Guat 5
## 19 Xin Oort 4
## 20 Kom <NA> 2## Number of NA values in MaleID: 0## Number of NA values in FemaleID: 0- After the use of the conditions in FemaleID I could see the changes where successfully done and that 0 NA’s are remaining in both FemaleID and MaleID
2.12 Dyad Response (7)
- The last variable I still have to treat for NA’s is DyadResponse. Before treating the NA’s we had 47 NA’s we know that we treated most of them we have only 7 remaining. These NA’s can be found at the rows 871, 1163, 1219, 1339, 1579,1888 and 1962
## Rows with missing values in DyadResponse: 871, 1163, 1219, 1339, 1579, 1888, 1962## Lines with missing values in DyadResponse:## # A tibble: 7 × 16
## Time Date Group MaleID FemaleID FemaleCorn DyadDistance
## <chr> <dttm> <chr> <chr> <chr> <dbl> <dbl>
## 1 09:39:26 2023-07-22 00:00:00 Ankhase Buk Ndaw 7 0
## 2 10:13:56 2023-06-23 00:00:00 Ankhase Sho Ginq 0 4
## 3 08:34:45 2023-06-17 00:00:00 Baie Dan… Kom Oort 3 2
## 4 08:54:12 2023-06-09 00:00:00 Baie Dan… Xia Piep 1 0
## 5 13:35:08 2023-05-03 00:00:00 Baie Dan… Kom Oort 5 3
## 6 13:27:30 2023-01-18 00:00:00 Ankhase Buk Ndaw 5 4
## 7 08:36:49 2022-12-13 00:00:00 Baie Dan… Kom Oort 8 4
## # ℹ 9 more variables: DyadResponse <chr>, OtherResponse <chr>, Audience <chr>,
## # IDIndividual1 <chr>, IntruderID <chr>, Remarks <chr>, MaleCorn <dbl>,
## # Intrusion <dbl>, AmountAudience <dbl>Row 871: The previous row was tolerance at 1m and the next tolerance at 0 which means that the row 871 should be Tolerance for DyadResponse
Row 1163: The value can not be found from the other rows so I will delete row 1163
Row 1219: The previous row was not approaching at 2m and the next is tolerance at 2m and tolerance at 1m, which means that the row 1219 should be Tolerance for DyadResponse
Row 1339: The previous row was tolerance at 0m while the next one was tolerance at 0m, which means that the tow 1339 should be Tolerance for DyadResponse
Row 1579: The value can not be found from the other rows so I will delete row 1579
Row 1888: The value can not be found from the other rows so I will delete row 1888
Row 1962: The value can not be found from the other rows so I will delete row 1962
## Number of remaining NA values in DyadResponse: 02.13 Final check: remaing NA’s in Bex?
## Final check of NA values in Bex:## Time Date Group MaleID FemaleID
## 0 0 0 0 0
## FemaleCorn DyadDistance DyadResponse OtherResponse Audience
## 0 0 0 0 0
## IDIndividual1 IntruderID Remarks MaleCorn Intrusion
## 0 0 0 0 0
## AmountAudience
## 03. Correction and creation of New Variables
3.1 Making a backup of Bex
- Before making new changes I will make a backup of my dataset at this point
3.2 Treating Remarks before processing with new modifications
Since I have removed all the missing data from the different columns, I now have to correct potential mistakes that can be found and create new variables to be able to manipulate better my data.
Since the column remarks contains corrections and additional information, I will treat it now
Before that lets check how many remarks we have in our dataset, how many of the main keywords we can find and make a visual representation of it
3.2.2 Vizualization of the Remarks Keywords
- Before making any changes I will make a count of the total amount of remarks and a count and barplot of the main keywords in the column to see in which proportion they are found. It has to be noted that some of the words are used in different contexts and have different meaning. This is why I will clean them manually
## Number of 'No Remarks' entries: 2139## Number of actual remarks entries: 603- There will be 599 entries I will have to treat manually in the Excel Spreadsheet for the Remarks
## Total number of keyword occurrences in the Barplot: 8223.2.3 Exporting of Bex
- I will now export the dataset and treat manually treat the remarks in an Excel spreadsheet before uploading it again and creating a new dataframe. I will also print a Glimpse of Bex to have information before the manual changes
## Glimpse of the Bex Before treating Remarks:## Rows: 2,742
## Columns: 16
## $ Time <chr> "09:47:50", "09:50:07", "09:53:11", "09:54:28", "09:55:…
## $ Date <dttm> 2022-09-27, 2022-09-27, 2022-09-27, 2022-09-27, 2022-0…
## $ Group <chr> "Baie Dankie", "Baie Dankie", "Baie Dankie", "Baie Dank…
## $ MaleID <chr> "Nge", "Nge", "Nge", "Nge", "Nge", "Nge", "Nge", "Nge",…
## $ FemaleID <chr> "Oerw", "Oerw", "Oerw", "Oerw", "Oerw", "Oerw", "Oerw",…
## $ FemaleCorn <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 7, 7, 7, 7…
## $ DyadDistance <dbl> 2, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 2, 1, 1, 0, 0…
## $ DyadResponse <chr> "Tolerance", "Tolerance", "Tolerance", "Tolerance", "To…
## $ OtherResponse <chr> "No Response", "No Response", "No Response", "No Respon…
## $ Audience <chr> "Obse; Oup; Sirk", "Obse; Oup; Sirk", "Oup; Sirk", "Sir…
## $ IDIndividual1 <chr> "No individual", "No individual", "No individual", "No …
## $ IntruderID <chr> "No Intrusion", "No Intrusion", "No Intrusion", "No Int…
## $ Remarks <chr> "No Remarks", "No Remarks", "Nge box did not open becau…
## $ MaleCorn <dbl> 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3…
## $ Intrusion <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0…
## $ AmountAudience <dbl> 3, 3, 2, 1, 2, 2, 2, 1, 1, 2, 6, 6, 3, 2, 2, 2, 2, 2, 2…## [1] "/Users/maki/Desktop/Master Thesis/BEX 2223 Master Thesis Maung Kyaw/IVPToleranceBex2223"3.2.4 Journal of manual changes in Bex excel spreadsheet
Before treating all the data in the Remarks I will create a few columns to redistribute information
- Context: To add contextual information
- SpecialBehaviour : To report any particular behaviour an individual could have done during a trial
- Got corn, to see if the individual got the corn or not
Also whenever i will have treated a remark, i will replace it with “Treated”. And if I have to delete the row I’ll write “Delete”. After re importing the data I will make a count of these changes to see if I still have the correct amount of cells and changes that have been done
- Creation of the columns Context, SpecialBehaviour and GotCorn
- Default values for the new columns are NoContext, NoSpecialBehaviour & Yes
a.Context: BoxMalfunction, BoxOpenedBefore, NoExperiment, Agonistic, Guat;Ap;Xian, CornLeak, BetweenGroupEncounter, ContactCalling,
b.SpecialBehaviour Oerw;Vo;Exp, Sey;Ap;AfterOpen, Oerw;Vo;Exp,Nge;Vo;Exp, Sirk;ApAfter30, Sirk;Av;Oerw, Oerw;Lo,Sey;Sf;Oort,Oort;At;Kom, Kom;Ap;AfterOpen, Sey;Ch,Sirk, Xin;Hesitation. Xia;Sf;Piep, Pom;Sf;Xian,Kom;Sf;Oort, Sey;Sf;Sirk, Xia;Sf;Piep,Piep;Sf,XIa, Oort;At;Kom, Sey;Rt;Sho;Ap, Sho;Rt;Ginq;Ap, Buk;Sf;Ndaw, Sho;Rt;Ndaw;Ap, Oort;Sf;Kom, Ginq;Sho;Ap;After30, Ndaw;Sc,Buk;Sf, Ndaw;Ap;After30, Kom;Ap;After30, Xia;Piep;Ap;After30, Pom;Bi;Xian, Sho;Ndaw;Av;Buk, Kom;Sf;Oort, Kom;St;Oort,Oort;St;Kom, Sey;Hi;Sirk, Obse;Ap;Piep;Av,Piep;Sf;Xia, Sirk;ApWhenPartnerLeft, Sey;Hh;Sirk, Xia;Sf;Piep;Sc, Xia;Piep;ShareFood, Piep;Ap;After30,Xia;Mu;Piep, Oort;St;Sirk;Ja,Sey;Sf, Pom;Sf;Xian, Ndaw;ApWhenPartnerLeft, Xian;At;Pom,Gaya;Su, Xian;Sf;Pom, Xian;Hesitation, Xia;ApWhenPartnerLeft,Sirk;Hesitation, Ginq;Hesitation, Sey;Ap;Kom;Av, Oort;Sc;Kom, Xian; Pom, Pom;Ap;Xian, Pom;Ap;Xian,Xian;Rt, Sey;Ap;Sirk;Rt, Sey;St;Sirk;Ig, Xia;Asf;Piep, Piep;ApWhenPartnerLeft, Sho;Ap;After30, Ginq;ApWhenPartnerLeft, Pom;Sf;Xian;Sf;Pom, Xian;ApWhenPartnerLeft, Piep;Ch;Sirk, Sey;St;Sirk, Ndaw;Ap;After30, Xian;Ap;After30, Xian;St;Prai, Pom;Sf;Xian;Vc, Kom;Ap;After30, Kom;ApproachWithPartner, Oort;ApWhenPartnerLeft, Sho;Ap;After30,Ginq;Ap;After30, Ginq;ApproachWithPartner, Ndaw;Hesitation, Oerw;Hesitation, Oerw;ApWhenPartnerLeft. Piep;Ap;After30, Sirk;Ap;After30, Xia;Ap;After30, Ouli;Gr;BBOuli, Oerw;Ap;After30, Sirk;Hesitation, Sey;Ap;Sirk;Av, Ouli;Ap;Xia;Av, Xin;Ap;After30, Sho;Sf;Ginq;Sc, Xia;ApWhenPartnerLeft, Sey;Ap;Sirk;Ja, Nge;Oerw;ShareFood, Nge;Ap;Oerw;Oerw;At,Obse;At;Nge,
c.GotCorn: No;Nge, No;Piep, No;Xian, No;Oort, No;Sirk, No;Kom, No;Ndaw, No;Kom, No;Oort, No;Xia, No;Buk, No;Sho, No;Sey,No;Piep, No;Ginq
- Remarks: Treated, TODelete
- Values set for exsting columns
IntruderID: Sey, Oerw, Guat, Kom, Gris, Sho, Oerw; Ouli, Guat; Gri, Xop, Obse, Oort, Obse; Sey, Ginq; Ghid, Xia, Grif, Sey, Gree; Gran, Godu; Gub, Gran, Oerw; Nak, Ghid, Buk, Oup
DyadDistance: 6, 7, 8, 9 , 1
Audience: UnidentifiedAudience, Ouli; Riss, Gris, Sey, Sey; Piep; Sirk, Oup Ome
IDIndividual1: Piep, Oort; Kom, Ndaw; Buk, Sho; Ginq, Ndaw, Buk, Xian, Pom, Oort; Kom, Buk; Ndaw, Sirk; Sey, Xin; Ouli, Oerw; Nge
DyadResponse: Tolerance, Not approaching; Losing interest, Losing interest; Intrusion
3.3 Re Uploading the dataset after the treatment of the Remarks
- Now that I have manually treated the remarks in a spreadsheet I will re import the dataset
## Rows: 2,742
## Columns: 19
## $ Time <chr> "09:47:50", "09:50:07", "09:53:11", "09:54:28", "09:5…
## $ Date <dttm> 2022-09-27, 2022-09-27, 2022-09-27, 2022-09-27, 2022…
## $ Group <chr> "Baie Dankie", "Baie Dankie", "Baie Dankie", "Baie Da…
## $ MaleID <chr> "Nge", "Nge", "Nge", "Nge", "Nge", "Nge", "Nge", "Nge…
## $ FemaleID <chr> "Oerw", "Oerw", "Oerw", "Oerw", "Oerw", "Oerw", "Oerw…
## $ FemaleCorn <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 7, 7, 7,…
## $ DyadDistance <dbl> 2, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 2, 1, 1, 0,…
## $ DyadResponse <chr> "Tolerance", "Tolerance", "Tolerance", "Tolerance", "…
## $ OtherResponse <chr> "No Response", "No Response", "No Response", "No Resp…
## $ Audience <chr> "Obse; Oup; Sirk", "Obse; Oup; Sirk", "Oup; Sirk", "S…
## $ IDIndividual1 <chr> "No individual", "No individual", "No individual", "N…
## $ IntruderID <chr> "No Intrusion", "No Intrusion", "No Intrusion", "No I…
## $ Remarks <chr> "No Remarks", "No Remarks", "Treated", "Treated", "No…
## $ MaleCorn <dbl> 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,…
## $ Intrusion <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0,…
## $ AmountAudience <dbl> 3, 3, 2, 1, 2, 2, 2, 1, 1, 2, 6, 6, 3, 2, 2, 2, 2, 2,…
## $ Context <chr> "NoContext", "NoContext", "BoxMalfunction", "BoxOpene…
## $ SpecialBehaviour <chr> "NoSpecialBehaviour", "NoSpecialBehaviour", "Oerw;Vo;…
## $ GotCorn <chr> "Yes", "Yes", "No;Nge", "Yes", "Yes", "Yes", "Yes", "…- And Check for any remaining NA’s in BexClean
## Number of NA entries in Context: 0## Number of NA entries in SpecialBehaviour: 0## Number of NA entries in GotCorn: 0## Number of NA entries in BexClean: 03.4 Time - Creation of Period and Hour
Time : I considered looking at the time sections in which we did the expermiment. I will thus look at the time ranges (max and min in the day / latest and earliest time) before separating the day in different sections to have an idea in which part of the day most of the experiments occured. This will not be used in my analysis, but if I wanted to, I could interesting to compare the amount of experimentations made per day and have a line indicating the time of sunrise.
The Minimum Time in the dataset is 06:03:26* while the Maximum Time is at 16:36:59
In my box experiment I have this variable called time that tells me when the experiment was done. I don’t think I need this information per se. I was wondering if it could be easy and interesting to see from when to when the time occurs and then separate this time in a few sections like early, monring, morning, miday, afternoon, end of the day
a.6 to 8 : Early morning b.8 to 10: Morning c.10 to 12: Noon d.12 to 14: Afternoon e.14 to 17: End of the day
Last, I want to create a variable called Hour that will take the value in Time and round it to the hour in which it is ex: from 06:00 to 06:59 -> 6, from 07:00 to 07:59 -> 7 etc…
This will allow me to see when most of the trials occured with more detail and I will be to see in which hour most of the trial happened. Nevertheless Period will be better for an improved readability
3.5 Date - Creation of Month and Day
- i want to Create a variable called month to see the month of the experiment and day so I know which day of the experiment it was (1st, 10th, 1000th..)
3.6 Male and Female ID - Creation of Dyad, Trial and Session
- I will use Female and Male ID to create different variables
- While checking if there are still any mistakes in FemaleID and MaleID using unique, I saw that some of the names are in the wrong rows. I want the 3 letter male codes whether they are in the column FemaleID or MaleID to be in the new column Male while I want the 4 letter female codes whether they are in FemaleID or MaleID to be in the new column Female before checking again with unique that the transformation worked. I will use mutate
## Unique Female IDs: Sirk Ginq Piep Oerw Xin Ndaw Xia Sey Ouli Nge Oort Xian Guat Kom## Unique Male IDs: Sey Sho Xia Nge Ouli Buk Piep Sirk Xin Oerw Kom Pom Oort| Buk | Kom | Nge | Oerw | Oort | Ouli | Piep | Pom | Sey | Sho | Sirk | Xia | Xin | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ginq | 6 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 277 | 0 | 0 | 0 |
| Guat | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 5 | 0 | 0 | 0 | 0 | 0 |
| Kom | 0 | 0 | 0 | 0 | 29 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Ndaw | 259 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Nge | 0 | 0 | 0 | 19 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Oerw | 0 | 0 | 164 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Oort | 0 | 337 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 4 |
| Ouli | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 159 |
| Piep | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 575 | 0 |
| Sey | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 17 | 0 | 0 |
| Sirk | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 570 | 0 | 0 | 0 | 0 |
| Xia | 0 | 0 | 0 | 0 | 0 | 0 | 35 | 0 | 0 | 0 | 0 | 0 | 0 |
| Xian | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 259 | 0 | 0 | 0 | 0 | 0 |
| Xin | 0 | 0 | 0 | 0 | 0 | 27 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
- Create a variable called Male that in each row will take the name of the 3 letter code that is either in MaleID or Female ID and a variable called Female that in each row will take the name of the 4 letter code that is either in MaleID or FemaleID
## Unique Dyads: Sey Sirk Sho Ginq Xia Piep Nge Oerw Xin Ouli Buk Ndaw Buk Ginq Kom Oort Pom Xian Pom Guat Xin Oort| Var1 | Freq |
|---|---|
| Buk Ginq | 6 |
| Buk Ndaw | 259 |
| Kom Oort | 366 |
| Nge Oerw | 183 |
| Pom Guat | 5 |
| Pom Xian | 259 |
| Sey Sirk | 587 |
| Sho Ginq | 277 |
| Xia Piep | 610 |
| Xin Oort | 4 |
| Xin Ouli | 186 |
## Unique Male-Female Combinations:## # A tibble: 11 × 2
## Male Female
## <chr> <chr>
## 1 Sey Sirk
## 2 Sho Ginq
## 3 Xia Piep
## 4 Nge Oerw
## 5 Xin Ouli
## 6 Buk Ndaw
## 7 Buk Ginq
## 8 Kom Oort
## 9 Pom Xian
## 10 Pom Guat
## 11 Xin Oort- Create the variable called Dyad created by combining the name of FemaleID and MaleID into one name with a space between the two codes. For information the 3 letter code is the name of the female while the 4 letter code is the name of the male like displayed here with the correct dyads;
- The correct dyads are:
- Buk Ndaw
- Kom Oort
- Nge Oerw
- Pom Guat
- Pom Xian
- Sey Sirk
- Sho Ginq
- Xia Piep
- Xin Ouli
- Buk Ndaw
- While the dyads we have now in the datset are;
- Buk Ginq 6
- Buk Ndaw 259
- Kom Oort 366
- Nge Oerw 183
- Pom Guat 5
- Pom Xian 259
- Sey Sirk 587
- Sho Ginq 277
- Xia Piep 610
- Xin Oort 4
- Xin Ouli 186
## [1] "Wrong Rows:"## [1] 613 614 615 616 617 931 2710 2711 2712 2713## [1] "Wrong Dyads:"## [1] "Buk Ginq" "Buk Ginq" "Buk Ginq" "Buk Ginq" "Buk Ginq" "Buk Ginq"
## [7] "Xin Oort" "Xin Oort" "Xin Oort" "Xin Oort"They are a 10 wrong dyads that I will have to identify in the dataset and manually correct, those wrong dyads to change and identify are: -Buk Ginq - 6 occurences -Xin Oort - 4 occurences
I will change the occurences of Buk Ginq to Sho Ginq for row 613 to 617 and row 931. I know these trials are with Sho Ginq because the comments mentioned Sho in them while Male(ID) gave Buk which was a mistake
For the rows from 2710 to 2713 since, Ouli is in the audience it is unlikely that we had trials with the dyad Xin Ouli. Also I think they are little chances that the names of both individuals were entered wrong. I will replace these occurences where we had Xin Oort by Kom Oort
I thus want Buk to be replaced in male ID in rows 613 to 617 and row 913 with Sho and, Xin to be replaced by Kom in rows 2710 to 2713 in Male ID before updating Dyad
If Rows 613 to 617 and 931 are coded with Buk for MaleId and Ginq for FemaleId replace Male by Sho
## Rows to correct Sho Ginq:## [1] 613 614 615 616 617 931## Rows to correct Kom Oort:## [1] 2710 2711 2712 2713## Wrong Rows After Correction:## integer(0)## Wrong Dyads After Correction:## character(0)## All dyads are now correct.## Unique Dyads after correction: Sey Sirk Sho Ginq Xia Piep Nge Oerw Xin Ouli Buk Ndaw Kom Oort Pom Xian Pom Guat| Var1 | Freq |
|---|---|
| Buk Ndaw | 259 |
| Kom Oort | 370 |
| Nge Oerw | 183 |
| Pom Guat | 5 |
| Pom Xian | 259 |
| Sey Sirk | 587 |
| Sho Ginq | 283 |
| Xia Piep | 610 |
| Xin Ouli | 186 |
## Number of rows changed to Sho Ginq: 6## Number of rows changed to Kom Oort: 4Create the variable called Trial where the data will be sorted by date and dyad in order to see how many trials have been done with each individual: One row (per dyad) = one trial and the variable called Day where the data will be sorted by date and dyad and day in order to see how many sessions have been done with each individual: One day (per dyad) = one session Now, let’s proceed with creating the Dyad variable, Trial, and Day:
Make a summary of trial and session so I can see see how many trials and sessions have been done with the individuals
## Trial Summary:| Dyad | Amount of Trials |
|---|---|
| Xia Piep | 610 |
| Sey Sirk | 587 |
| Kom Oort | 370 |
| Sho Ginq | 283 |
| Buk Ndaw | 259 |
| Pom Xian | 259 |
| Xin Ouli | 186 |
| Nge Oerw | 183 |
| Pom Guat | 5 |
##
## Day Summary:| Dyad | Number of Days |
|---|---|
| Xin Ouli | 276 |
| Xia Piep | 249 |
| Sho Ginq | 200 |
| Sey Sirk | 165 |
| Pom Xian | 112 |
| Pom Guat | 93 |
| Nge Oerw | 92 |
| Kom Oort | 70 |
| Buk Ndaw | 39 |
- After relfection I decided to remove every column that is with PomGuat since they are not enough trials for this Dyad and since we then changed PomGuat for PomXian. I have 5 occurnces to change. For easier manipulation I will remove every row where there is Guat
## Change in Rows: -53.7 Female and Male Placement Corn
- The idea is that for each Dyad, we gave an amount of corn to attract the monkey of a dyad to the right distance of his partner for a trial by putting corn in experiment box that he could get by approaching. We repeated this step as much as needed to have our Dyad at the desired distance to continue the trials from the previous day of experimentation. This means I will only Keep the last number per dyad and day for each day.
- I decided to create two variables, that are called PlacementMale and PlacementFemale that will only keep the final amount of corn given to each individual within a day of experiment
## Placement columns were created successfully.
3.8 DyadDistance - Creation of proximity
- I would like to create a variable called proximity to have another measure of the proximity of the individuals.
- First lets look at the maximum and minimum distance found in Dyad Distance
## Maximum Distance: 10## Minimum Distance: 0- The minumum Distance is 0m while the maximum Distance is 10m
- Then lets make a graph of the frequency for each distances with a line for the median to show the most frquent distance
- Now lets create a new variable called Proximity using the distances found in DyadDistance on the following model:
- 0 = Contact
- 1 - 2 = 1-2
- 2 - 3 = 2-3
- 4 - 5 = 4-5
- 5 - 10 = +5
3.9.1 Dyad Response - Corrections and detailed cleaning
Reminder: The different behaviors that are coded in DyadResponse are: Distracted, Female aggress male, Male aggress female, Intrusion, Loosing interest, Not approaching, Tolerance and Other
- I will change the columns associated to each behavior (i.e. Response) of DyadResponse into dichotomic variables in order to see the frequency of each behaviour
- This will allow me to see which behavior occurred more than others, and what differences are between dyads
- As multiple behaviours could occur within the same trial, multiple responses (data entries) can be found in a single cell. I will create a hierarchy to reduce the amount of behaviors assigned to each trial (if there is more than one).
- correct any potential discrepancies (ex. if tolerance and aggression occured within the same trial aggression>tolerance)
- assign as few labels per trial, ideally one using a hierarchy among the occuring behaviours to choose which response to keep
- get a better view and understanding of the data and the most common behaviours produced by each dyad producing plots and tables
- if necessary create new variables that can complement redistribute the information initally found in the column
I will create some tables to have a better understanding of the state of the column dyadresponse and the different existing combinations at this point
Also I will create the hierarchy before implementing in the dataset
3.9.2 Dyad Response description
- I need to know how many different combinations exist within DyadResponse and how many occurnces/cells have more than one response per cell. I need the code to not take in account the order but rather the responses in themselves
- I want one table for the Different combinantions and one showing the differernt combinations for only ONE response per cell, in a second table
## Number of rows with multiple entries in DyadResponse: 271## Rows with multiple entries in DyadResponse: 3 4 13 53 58 64 68 71 72 84 88 89 97 98 113 133 137 192 194 244 275 277 278 284 289 298 302 304 318 322 326 327 328 331 341 346 347 367 368 387 388 391 397 409 454 468 469 482 497 510 511 522 529 536 597 600 626 629 687 696 706 711 713 726 740 748 749 757 760 761 763 768 769 780 786 818 829 841 843 845 861 866 877 888 892 898 912 919 934 937 938 954 955 962 973 997 1004 1015 1027 1042 1048 1060 1108 1141 1143 1161 1162 1190 1209 1213 1217 1225 1229 1236 1240 1243 1245 1247 1254 1267 1271 1284 1288 1293 1308 1310 1311 1326 1327 1330 1336 1341 1343 1352 1378 1398 1399 1407 1416 1461 1464 1473 1478 1484 1488 1489 1495 1510 1520 1521 1528 1558 1595 1598 1602 1607 1616 1629 1630 1633 1636 1639 1642 1645 1654 1658 1679 1708 1714 1718 1724 1741 1750 1752 1784 1785 1788 1795 1800 1801 1802 1804 1805 1806 1807 1808 1815 1816 1819 1820 1821 1822 1824 1825 1828 1829 1830 1831 1832 1837 1843 1856 1857 1858 1889 1897 1968 1984 2023 2066 2088 2090 2091 2100 2101 2102 2103 2104 2105 2109 2112 2113 2147 2148 2149 2152 2153 2166 2175 2184 2185 2192 2193 2194 2199 2201 2202 2205 2218 2226 2237 2248 2257 2290 2294 2343 2353 2354 2359 2360 2362 2363 2370 2388 2395 2396 2400 2465 2466 2467 2493 2534 2601 2609 2626 2629 2630 2643 2651 2657 2730## Total number of rows in DyadResponse: 2737## Number of rows with a single entry in DyadResponse: 2466## Number of rows with multiple entries in DyadResponse: 271## Sum of single and multiple entry rows: 2737##
##
## Table: Unique DyadResponse Combinations and Their Frequencies
##
## |Combination | Frequency|
## |:-------------------------------------------------|---------:|
## |Intrusion;Tolerance | 49|
## |Losing interest;Not approaching | 47|
## |Intrusion;Not approaching | 30|
## |Male aggress female;Tolerance | 27|
## |Looks at partner;Tolerance | 23|
## |Losing interest;Tolerance | 22|
## |Female aggress male;Tolerance | 19|
## |Looks at partner;Not approaching | 9|
## |Other;Tolerance | 7|
## |Distracted;Not approaching | 6|
## |Not approaching;Tolerance | 4|
## |Distracted;Losing interest | 3|
## |Female aggress male;Male aggress female;Tolerance | 3|
## |Female aggress male;Not approaching | 3|
## |Distracted;Tolerance | 2|
## |Female aggress male;Intrusion;Tolerance | 2|
## |Male aggress female;Not approaching | 2|
## |Distracted;Intrusion;Not approaching | 1|
## |Distracted;Intrusion;Tolerance | 1|
## |Female aggress male;Intrusion | 1|
## |Intrusion;Losing interest | 1|
## |Intrusion;Losing interest;Not approaching | 1|
## |Intrusion;Male aggress female | 1|
## |Intrusion;Not approaching;Tolerance | 1|
## |Looks at partner;Losing interest;Not approaching | 1|
## |Looks at partner;Male aggress female | 1|
## |Looks at partner;Male aggress female;Tolerance | 1|
## |Looks at partner;Other;Tolerance | 1|
## |Losing interest;Not approaching;Tolerance | 1|
## |Not approaching;Other | 1|##
##
## Table: Rows with Single Responses in DyadResponse
##
## |DyadResponse | Frequency|
## |:-------------------|---------:|
## |Tolerance | 1809|
## |Not approaching | 465|
## |Male aggress female | 73|
## |Intrusion | 51|
## |Losing interest | 37|
## |Female aggress male | 18|
## |Other | 9|
## |Distracted | 4|##
##
## Table: Rows with Multiple Responses in DyadResponse
##
## |Combination | Frequency|
## |:-------------------------------------------------|---------:|
## |Intrusion;Tolerance | 49|
## |Losing interest;Not approaching | 47|
## |Intrusion;Not approaching | 30|
## |Male aggress female;Tolerance | 27|
## |Looks at partner;Tolerance | 23|
## |Losing interest;Tolerance | 22|
## |Female aggress male;Tolerance | 19|
## |Looks at partner;Not approaching | 9|
## |Other;Tolerance | 7|
## |Distracted;Not approaching | 6|
## |Not approaching;Tolerance | 4|
## |Distracted;Losing interest | 3|
## |Female aggress male;Male aggress female;Tolerance | 3|
## |Female aggress male;Not approaching | 3|
## |Distracted;Tolerance | 2|
## |Female aggress male;Intrusion;Tolerance | 2|
## |Male aggress female;Not approaching | 2|
## |Distracted;Intrusion;Not approaching | 1|
## |Distracted;Intrusion;Tolerance | 1|
## |Female aggress male;Intrusion | 1|
## |Intrusion;Losing interest | 1|
## |Intrusion;Losing interest;Not approaching | 1|
## |Intrusion;Male aggress female | 1|
## |Intrusion;Not approaching;Tolerance | 1|
## |Looks at partner;Losing interest;Not approaching | 1|
## |Looks at partner;Male aggress female | 1|
## |Looks at partner;Male aggress female;Tolerance | 1|
## |Looks at partner;Other;Tolerance | 1|
## |Losing interest;Not approaching;Tolerance | 1|
## |Not approaching;Other | 1|
### 3.9.3 Dyad Response Hierarchy
Dyad Response Hierarchy Projection of the hierarchy (changes will be made) - Aggression > Tolerance - Tolerance > Not approaching -> Create a variable called hesistant in addtion to the tolerance count to see frequency of tolerance behaviour that happened after > 1min - Tolerance > Loosing interest - Tolerance > Intrusion
- Not approaching = looking box but not coming while Loosing interest = not paying attention to the box - Intrusion > Loosing interest - Intrusion > Not approaching - Not approaching > Looks at partner - We can code every look at partner as no approaching and keep the count of looks at partner as additional information
- Not approaching >?> Loosing interest ? !! - Define distracted - Not approaching > Distracted - Aggression > Not approaching - Other > Look case by case and categorize depending of behavior
- Remarks may be used for the same reasonFirst I want to see how many rows in DyadResponse have more than one entry per cell
- And if there are more than one value per cell, report it in a new column called “MultipleResponses”
Now that I know that they are 230 rows with multiple entries I will print each combinations to see which one I have to print and also display the combinations once the data is split.
I will make an intermediary summary to see that state of DyadResponse at the moment
A. Summary od DyadResponse
B.Idnetify Rows with more than 1 entry
Step 1: Identify Rows with More than 1 Entry beforesummaryrowswithmultipleentries_1 <- which(sapply(BexClean$DyadResponse, function(x) length(unlist(strsplit(as.character(x), “;”))) > 1))
Display the rows with more than 1 entry in DyadResponse knitr::kable(BexClean[beforesummaryrowswithmultipleentries_1, ])
Print the amount of cells with more than 1 entry and the total number of rows cat(“Number of rows with more than 1 entry in DyadResponse:”, length(beforesummaryrowswithmultipleentries_1), “”)
C. Identify Rows with more than 2 entry
## Number of rows with more than 1 entry in DyadResponse: 271## Unique Combinations and Counts for More than 1 Entry:## Distracted & Intrusion 2
## Female aggress male & Intrusion 2
## Looks at partner & Other 1
## Losing interest & Intrusion 1
## Losing interest & Looks at partner 1
## Male aggress female & Female aggress male 3
## Male aggress female & Looks at partner 1
## Not approaching & Intrusion 1
## Not approaching & Losing interest 1
## Distracted & Losing interest 3
## Female aggress male & Intrusion 1
## Female aggress male & Not approaching 3
## Losing interest & Intrusion 1
## Male aggress female & Intrusion 1
## Male aggress female & Looks at partner 1
## Male aggress female & Not approaching 2
## Not approaching & Distracted 7
## Not approaching & Intrusion 32
## Not approaching & Looks at partner 10
## Not approaching & Losing interest 49
## Not approaching & Other 1
## Tolerance & Distracted 3
## Tolerance & Female aggress male 24
## Tolerance & Intrusion 53
## Tolerance & Looks at partner 25
## Tolerance & Losing interest 23
## Tolerance & Male aggress female 31
## Tolerance & Not approaching 6
## Tolerance & Other 8## Change in Occurrences for the Chunk (Female aggress male > Not approaching): 2731 2737 2737 2478 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2400 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2573 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2718 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2708 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2710 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2702 2737 2737 2737 2732 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2478 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2167 2737 2737 2737 2460 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2720 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2162 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2737 2733 2578 2737 2737 2737 2737 2737 2737- Male agress female > Not approaching : Remove Not approaching if there is Male aggress female (2x)
## Change in Occurrences for the Chunk (Male aggress female > Not approaching): -2- Tolerance > Distracted: Remove Distracted if there is tolerance (3x)
## Change in Occurrences for the Chunk Tolerance > Distracted: -38.Female agress male > Tolerance: Remove Tolerance if there is Female agress male (23x)
## Change in Occurrences for the Chunk (Tolerance & Female aggress male): -249.Tolerance > Loosing interest: Remove Losing interest if there is Tolerance (13x)
## Change in Occurrences for the Chunk (Tolerance & Losing interest): 010.Male agress female > Tolerance : Remove Tolerance if there is Male agress male (31)
## Change in Occurrences for the Chunk (Tolerance & Male aggress female): -2811.Tolerance > Not approaching: Remove Not approaching if there is tolerance (6x)
## Change in Occurrences for teh Chunk (Tolerance & Not approaching): -6- Looks at partner & Other 1 - Display line for detailed check (1x)
- Not approaching & Losing interest 46 - Keep both, I will consider doing a detailled check
- Not approaching & Distracted 7 - Display line for detailed check
- Not approaching & Other 1 - Display line for detailed check
- Tolerance & Other 8 - Display line for detailed check
SEE WHEN AND HOW TO Put ThIS CODE TO REPLACE PREIVOSU NUMBERS:
1. Aggression > Tolerance
BexClean <- BexClean[!(grepl(“Tolerance”, BexClean\(DyadResponse) & (grepl("Female aggress male", BexClean\)DyadResponse) | grepl(“Male aggress female”, BexClean$DyadResponse))), ]
2. Tolerance > Not approaching
BexClean <- BexClean[!(grepl(“Not approaching”, BexClean\(DyadResponse) & grepl("Tolerance", BexClean\)DyadResponse)), ]
3. Tolerance > Loosing interest
BexClean <- BexClean[!(grepl(“Losing interest”, BexClean\(DyadResponse) & grepl("Tolerance", BexClean\)DyadResponse)), ]
4. Tolerance > Intrusion
BexClean <- BexClean[!(grepl(“Intrusion”, BexClean\(DyadResponse) & grepl("Tolerance", BexClean\)DyadResponse)), ]
5. Intrusion > Loosing interest
BexClean <- BexClean[!(grepl(“Losing interest”, BexClean\(DyadResponse) & grepl("Intrusion", BexClean\)DyadResponse)), ]
6. Intrusion > Not approaching
BexClean <- BexClean[!(grepl(“Not approaching”, BexClean\(DyadResponse) & grepl("Intrusion", BexClean\)DyadResponse)), ]
7. Not approaching > Looks at partner
Code every look at partner as not approaching and keep the count of looks at partner as additional information
8. Define distracted
Not approaching > Distracted
BexClean <- BexClean[!(grepl(“Distracted”, BexClean\(DyadResponse) & grepl("Not approaching", BexClean\)DyadResponse)), ]
9. Other > Look case by case and categorize depending on behavior
Final clean-up and reporting
knitr::kable(BexClean, col.names = colnames(BexClean), caption = “Cleaned DyadResponse Data”)
Code to display rows numbers and the response in DyadResponse that have mroe than one entry
Other Reponse - DEtailed cleaning to delete the column
Audience - Creation of Amount Audience and Density
ID Individua1 - Not sure yet
Intruder ID
Remarks - Detailed Cleaning
Intrusion
Not Approaching
Lossing Interest
Distracted
MultipleResponse
Amount Audience
DyadDistance
Distance
No trial
Code to keep and place again
Code to settle
I also chose to directly create new dichotomic variables for “Not approaching”, “Intrusion”, “Losing interest”, “Distracted”, for this I would like the function to 1.Check if there is a value different than “No individual” 2.If the value ≠“No individual” then I want it to take the response found in “DyadResposne”
## The 'MultipleResponses' column already exists. No changes made.## Occurrences of 'Not approaching' in DyadResponse: 564## Occurrences of 'Intrusion' in DyadResponse: 135## Occurrences of 'Losing interest' in DyadResponse: 112## Occurrences of 'Distracted' in DyadResponse: 14## Warning: Unknown or uninitialised column: `NotApproaching`.## Warning: Unknown or uninitialised column: `LosingInterest`.## Warning: Unknown or uninitialised column: `Distracted`.## Occurrences of '1' in 'NotApproaching': 0## Occurrences of '1' in 'Intrusion': 56## Occurrences of '1' in 'LosingInterest': 0## Occurrences of '1' in 'Distracted': 0Exploratory Graph (To organise)
####Dyad, Distance & Date
Trials of grpahs, I will have to check all of them
My goal here is too see if each dyad have an general evolution of their dyad distance trough time and how many varaition do they have
## `geom_smooth()` using formula = 'y ~ x'
## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.## `geom_smooth()` using formula = 'y ~ x'
## `summarise()` has grouped output by 'Dyad'. You can override using the
## `.groups` argument.
- I want to use the date to know how many sessions have been done with each dyads in my experiment. * I will create a variable called Session where 1 session = 1 day * The data has values from the 14th of September 2022 until the 13th of September 2023 * I will also create a variable called Trial to know how many trials have been done with each dyad where 1 row = 1 trial
* I may consider, in parallel of my hypothesis, to separate the data in *4 seasons* to make a preliminary check of a potential effect of seasonality. Nevertheless the fact that we did not use anywithout tools to mesure the weather and the idea to make a categorization in 4 seasons without considering the actua quite arbitrary. I may do it but with no intention to include this in my scientific report.
l temperature, food quantitiy and other elements related to seasonailty make this categorizationn a categorization where 12 months of data will be separated in 4 categories- But before I may want to make a few changes already by merging Male corn and Male placement corn into ” Male corn” and maybe replacing all of the NA’s in “Other response” by response
#Lines to check unique values in MaleFemaleID to see if they are any problems with it # Unique values in MaleID unique_male_ids <- unique(BexClean$MaleID)
Unique values in FemaleID
unique_female_ids <- unique(Bex$FemaleID)
Sections below are here for the organization of my paper and will be worked on once the data cleaning and exploration is done
3. Describing the data
4.Visualizing the data
5.Research question & Hypothesis
Research question
What factors influence the rate at which individuals (vervets) learn to tolerate each other in a controlled box experiment?
Ex: The rate at which individuals (vervets) learn to tolerate each other in a box experiment is influenced by social factors (audience, social network, behavior of the partner) and idioyncratic factors (age, rank)
Hypothesis
- Hypothesis about the Presence of High-Ranking Individuals:
The presence of a higher number of high-ranking individuals in the audience will negatively correlate with the level of tolerance achieved among vervets in the box experiment. This is expected to result in higher frequencies of aggressive behaviors, intrusions, and loss of interest, particularly from lower-ranking individuals.
- Hypothesis about Partner Agonistic Behaviors:
Vervets tolerance levels in the box experiment will be influenced by their partner’s display of agonistic behaviors. Specifically, partners who exhibit more frequent agonistic behaviors towards their partner will lead to decrease in their motivation to participate in future trials.
- Hypothesis about the Establishment of an Optimal Distance:
During the box experiment, vervet dyads will establish an “optimal” distance for interaction, characterized by a higher frequency of tolerance compared to other distances. This optimal distance is expected to signify that the individuals tolerate each other more effectively at this specific proximity .
- Hypothesis about Age and Rank:
The age and rank of individual vervets within the group will influence the success of the trials in the box experiment. Specifically, older and higher-ranking individuals are expected to exhibit lower rates of success compared to dyads consisting of younger and lower-ranked individuals. This decrease in success is anticipated to be associated with a higher frequency of aggressive behaviors displayed by older and higher-ranking individuals towards their partners. (I’m not sure this hypothesis makes sens, I have the feeling age and rank must have an influence but I don’t know how to put it, I will think about it)
- Hyptohesis about seasonality
Seasonality is expected to impact the motivation of vervet dyads to participate in the box experiment. We hypothesize that dyads will have lower motivation, as indicated by a reduced number of trials, during the summer months compared to the winter months. This difference in motivation is likely influenced by temperature and food availability. To test this hypothesis, we will categorize the data into four seasonal periods, each spanning four months, and analyze whether there is a significant effect of seasonality on the motivation to engage in the trials.
6.Statistical tests and analisis of the data
Statistical tests
- Hypothesis 1: Influence of High-Ranking Individuals
Variables Needed:
DyadResponse (specifically, “aggression” responses) Amountaudience (to measure the number of individuals in the audience) Audience…15 (to identify the names of individuals in the audience for calculating dominance ranks) Elo rating of the individuals based on the ab libitum data collected in IVP (which I have to calculate asap)
Statistical Analysis:Logistic Regression, as it could analyze the influence of high-ranking individuals on the occurrence of aggression in dyad responses. This will help determine whether the presence of high-ranking individuals affects the likelihood of aggression.
- Hypothesis 2: Impact of Partner’s Agonistic Behaviors
Variables Needed:
- DyadResponse (specifically, “aggression” responses)
- MaleagressF (male’s aggression towards female)
- FemaleaggressM (female’s aggression towards male)
Statistical Analysis: Logistic Regression as it could be used to assess how the occurrence of aggression in dyad responses is influenced by the partner’s gender-specific agonistic behaviors.
- Hypothesis 3: Identification of an Optimal Interaction Distance
Variables Needed:
- DyadDistance (distance between boxes)
- Tolerance (as a binary outcome)
Statistical Analysis: generalized Linear Model (GLM) to investigate whether there is an optimal distance that leads to a higher likelihood of tolerance (Tolerance = 1).
- Hypothesis 4: Role of Age and Rank
Variables Needed:
- Tolerance (as a binary outcome)
- Male and Female (to identify individuals’ ages and ranks)
- Dyad (to link individuals to dyads)
- Birthdate to calculate the age of each individual
Statistical Analysis: Logistic Regression Logistic regression can be employed to determine whether the age and rank of individual vervets within dyads have an impact on the likelihood of tolerance (Tolerance = 1).
- Hypothesis 5: Influence of Seasonality
Variables Needed:
- Date (to categorize data into seasons)
- Trial (to count the number of trials in each season) and the data for at least 365 days so i can separate the data in 4 (1 year = 4 seasons = 12*4 month) to see if they may be an effect of seasonality on the motivation (amount of trials) of the dyads
Statistical Analysis:
ANOVA or Kruskal-Wallis Test: Depending on the distribution of your trial data, you can use either ANOVA (if the data are normally distributed) or the Kruskal-Wallis test (for non-normally distributed data) to assess the impact of seasonality on the number of trials. If significant differences are found, you can follow up with post-hoc tests to identify which seasons differ from each other. Please note that the effectiveness of these analyses may depend on the distribution of your data and specific research objectives. You may also consider conducting exploratory data analysis (e.g., visualization) to gain a better understanding of your dataset before performing these analyses. Additionally, if you have specific questions about data preprocessing or variable transformations, feel free to ask for further guidance. –> I took this from ChatGPT, I have to look more into it
REMARKS: So here are a few updates I made in the document. I also planned to send my cleaned data to Radu (the statistician of UNINE) as he was keen to help me find the right test. Of course I will also look again in Bshary’s and Charlotte’s work with the boxes and improve these suggestions that are quite simple for now
Also I still have to clean the last grpahs about male/female aggression as I didn’t finish that yet. I juste wanted to share my hypothesis and ideas for statistics so I can soon go into the “serious” work
Anyway, thank you in advance for your help <3
Michael
7. Plotting the results of the analysis
9. Interpretation of the results
10. Comeback on the research question and hypothesis
11. Bibliography
12. Organization for my paper
- Introduction
- Tolerance humans, primates
- Apes vs monkeys / Captivity vs Wild
- IVP: Wild habituated vervets, experiments possible
- Paper Bshary, Canteloup… Prolongation study
- Relevance idea/topic research
- Research question & hypothesis
But: intro need triangle shape: broad to narrow end wiht research question> tolerance importance > animal reign, actual knowledge/ direction knowledge we need > show how my experiment goes in that way How to adress the gap, answer with research question
Then explain why choosing vervet monkeys, (IVP in methods), sociality, experiments made
Methods
IVP, research area, (goal, house, type people)
Population: groups, dyads, male/female, ranks..
Box material: boxes, remotes, batteries, camera, tripod, corn (no marmelade ;), (water spray, security reason, non agressive way to select individuals and not engage with mokeys when reachrging boxes with corn), pattern, previous distances, tablets, box experiment form
Tablets
(No observers mentionned)
Habituation boxes > individuals trained to recognice boxes, they have differernt levels of habituation
Patterns > appendix, mention similar to habituation, use to recognize box but efficieny depeds of experience)
Selection dyads > assigment from elo rating (different rank), if above average bond no dyad made, if not possible, availibilty of monkey also factor !! Non random can be a problem, think about why and how you selected data We created variations in dyads made by different sex, rank and not above average bonde (calculate bondeness)
Amount corn, do you want to mention it> maybe important Calculate corn during and placement cf paper on corn /food motivation
Corn (daily intake vervet % made from corn, cf site we saw, cf screenshot, comapre paper previousely made an all)
1st dyad trial (BD) > appendix
Videos > details appendix
Finding dyads > appendix
Placement to attract them > meniton if statiscial made on placement corn
Trials (1 session = max 15 trials/in total) (session could be broken in different sub sessions to reach 15 trials max)
If agression > 1m / If 2x tolerance < 1m , also if not approaching > 1m ( if no tolerance increase distance except if intrusion) (borgeaud > expectation fo aggression)
Time of the day > appendix
Territory? > appendix
Amount sessions p day/week, how we chose the moment to follow them >appendix
Problems/ unplanned events: weather, BGE’s, not finding the monkeys (group, dyad or individual), dispersal of males, river crossing, inacessibility (experiments or boxes), low vision (experiments or monkeys),> appendix
(Where do i mention the confounding variables?) > look in litterature, if something that could affect and already reported in papers check, oterhwise exclude “normal life” factors for both monekys and Experimenter
Types of experimental plan
Statistical tests (for each hypothesis)
Analysis
Results
Interpretation
Conclusion
Glossary
- Tolerance: Tolerance: An individual has an encounter with a conspecific and can freely leave but remains in the encounter without acting aggressively toward the conspecific. (Pisor & Surbeck, 2019)
- Agression
- Session
- Trial
- Group: In the Primate order, groups are individuals “which remain [physically] together in or separate from a larger unit” and interact with each other more than with other individuals.6 This definition does not cover all uses of the word “group” in the social sciences (e.g., human identity groups who identify with a common name or symbol may or may not interact with one another more frequently than with other individuals). Because of this ambiguity, we use the word “community” when referring to humans to better capture the notion of spatial proximity, per Ref. 54. Members of the same group are referred to as “same-group” and those from another group “extra-group.” (Pisor & Surbeck, 2019)
Bibliography
• Pisor, A. C., & Surbeck, M. (2019). The evolution of intergroup tolerance in nonhuman primates and humans. Evolutionary Anthropology: Issues and ReViews. Advance online publication. https://doi.org/10.1002/evan.21793 (Pisor & Surbeck, 2019)
Annex
Annex 1 : View of the dataset when imported - First 6 entries of each variable
- We can see here the brief View of the original dataset names BoxExwhen i initially imported it as seen in section 0: Opening data
| Date | Time | Data | Group |
|---|---|---|---|
| 2022-09-27 | 1899-12-31 09:47:50 | Box Experiment | Baie Dankie |
| 2022-09-27 | 1899-12-31 09:50:07 | Box Experiment | Baie Dankie |
| 2022-09-27 | 1899-12-31 09:53:11 | Box Experiment | Baie Dankie |
| 2022-09-27 | 1899-12-31 09:54:28 | Box Experiment | Baie Dankie |
| 2022-09-27 | 1899-12-31 09:55:19 | Box Experiment | Baie Dankie |
| 2022-09-27 | 1899-12-31 09:56:56 | Box Experiment | Baie Dankie |
| GPSS | GPSE | MaleID | FemaleID |
|---|---|---|---|
| -28.010549999999999 | 31.191050000000001 | Nge | Oerw |
| -28.010549999999999 | 31.191050000000001 | Nge | Oerw |
| -28.010549999999999 | 31.191050000000001 | Nge | Oerw |
| -28.010549999999999 | 31.191050000000001 | Nge | Oerw |
| -28.010549999999999 | 31.191050000000001 | Nge | Oerw |
| -28.010549999999999 | 31.191050000000001 | Nge | Oerw |
| Male placement corn | MaleCorn | FemaleCorn | DyadDistance | DyadResponse |
|---|---|---|---|---|
| NA | 3 | NA | 2m | Tolerance |
| NA | 3 | NA | 2m | Tolerance |
| NA | 3 | NA | 1m | Tolerance |
| NA | 3 | NA | 1m | Tolerance |
| NA | 3 | NA | 0m | Tolerance |
| NA | 3 | NA | 0m | Tolerance |
| OtherResponse | Audience | IDIndividual1 | IntruderID |
|---|---|---|---|
| NA | Obse; Oup; Sirk | NA | NA |
| NA | Obse; Oup; Sirk | NA | NA |
| NA | Oup; Sirk | NA | NA |
| NA | Sirk | NA | NA |
| NA | Sey; Sirk | NA | NA |
| NA | Sey; Sirk | NA | NA |
| Remarks |
|---|
| NA |
| NA |
| Nge box did not open because of the battery. Oerw vocalized to MA when he ap to the box to open it. |
| Sey came to the boxes once they were open |
| NA |
| NA |
| Observers | DeviceId |
|---|---|
| Josefien; Michael; Ona; Zonke | {7A4E6639-7387-7648-88EC-7FD27A0F258A} |
| Josefien; Michael; Ona; Zonke | {7A4E6639-7387-7648-88EC-7FD27A0F258A} |
| Josefien; Michael; Ona; Zonke | {7A4E6639-7387-7648-88EC-7FD27A0F258A} |
| Josefien; Michael; Ona; Zonke | {7A4E6639-7387-7648-88EC-7FD27A0F258A} |
| Josefien; Michael; Ona; Zonke | {7A4E6639-7387-7648-88EC-7FD27A0F258A} |
| Josefien; Michael; Ona; Zonke | {7A4E6639-7387-7648-88EC-7FD27A0F258A} |
```beforeNA